Japanese Localization and Input Support suggestions wanted

[Belxjander]

Other than the .language and .Charset files for locale.library... what else would be needed for Japanese to become more functional as a language on AmigaOS?

I'm also looking at what options there are for Japanese as an Input language option as well using a custom library and commodity setup...
any suggestions where I can avoid patching the OS would be appreciated.

-- EDIT --

I'm going with UTF-8 and ISO-2022-JP with ISO-2022-KR and ISO-2022-CN for the three language embedded locales for Japanese, Korean and Chinese,
adding UTF-8, UTF-16BE, UTF-16LE and UTF-7 outside that with a few other CJK locales just to cover bases...

Beyond the basic mapping requirements of encoding... I am definitely going to make use of codesets.library as one part of the backend and will enjoy seeing Japanese properly on-screen in AmIRC and WookieChat at least ... (try irc.juggler.jp 6667 sometime to get a sample)

if UTF-8 is already implimented in AmigaOS 4.x is it the default locale? and do I just need a font capable of properly mapping the characters concerned?

I'm also choosing to directly embed a base "perception" commodity directly inside the main library to enable some basic keyboard filter handling using commodities to allow for certain mode-changes to be processed properly (Hiragana/Katakana/Kanji/Han/... pictographic modes based on various Input styles)

Romaji / Pinyin / ... will be available once the language files register and start using perception.library more properly...

Do the coders for codesets.library come onto this board as well?

[chris]

I think the .charset support in OS4 only supports pure 8-bit character sets. Fonts and various catalog files are converted from UTF-8 to local charset - so there is some sort of UTF-8 support, but you can't directly print or use UTF-8 with OS functions.

[Belxjander]

I think the .charset support in OS4 only supports pure 8-bit character sets. Fonts and various catalog files are converted from UTF-8 to local charset - so there is some sort of UTF-8 support, but you can't directly print or use UTF-8 with OS functions.

well THAT would be a project killer... at any rate the way the Chinese, Japanese and Korean languages are encoded generally requires at least 2 octets (16bits) per character at the minimum.

It will be a serious pain if I have to sort out contacting and working with the locale.library author... and the example information generally lists up to 32bits per character which is the UCS2 and UCS4 standard size ... I was hoping to get the ISO-2022-* series of locale charsets setup at least.

I'm still working out the library specifics anyway... just committed a large resource change for some Linux IME material that I snapshot before it gets lost

[chris]

I think the .charset support in OS4 only supports pure 8-bit character sets. Fonts and various catalog files are converted from UTF-8 to local charset - so there is some sort of UTF-8 support, but you can't directly print or use UTF-8 with OS functions.

well THAT would be a project killer... at any rate the way the Chinese, Japanese and Korean languages are encoded generally requires at least 2 octets (16bits) per character at the minimum.

It will be a serious pain if I have to sort out contacting and working with the locale.library author... and the example information generally lists up to 32bits per character which is the UCS2 and UCS4 standard size ... I was hoping to get the ISO-2022-* series of locale charsets setup at least.

Well, you can select UTF-8 as a charset for windows and things, but the result is a crash, so I'm not sure where the problem lies.

[Belxjander]

I think the .charset support in OS4 only supports pure 8-bit character sets. Fonts and various catalog files are converted from UTF-8 to local charset - so there is some sort of UTF-8 support, but you can't directly print or use UTF-8 with OS functions.

well THAT would be a project killer... at any rate the way the Chinese, Japanese and Korean languages are encoded generally requires at least 2 octets (16bits) per character at the minimum.

It will be a serious pain if I have to sort out contacting and working with the locale.library author... and the example information generally lists up to 32bits per character which is the UCS2 and UCS4 standard size ... I was hoping to get the ISO-2022-* series of locale charsets setup at least.

Well, you can select UTF-8 as a charset for windows and things, but the result is a crash, so I'm not sure where the problem lies.

I mostly need the UTF-8 for remapping ISO-2022-JP, ISO-2022-JP-2, ISO-2022-CN, ISO-2022-CN-EXT, ISO-2022-KR, EUC-JP and EUC-KR charsets in Perception-IME,
starting with the ISO-2022-[JP/CN/KR] triplet first (in that order) for the .language file embedded charsets.